How to Write a Case Study That Closes

A case study closes when it leads with a specific number and a real human face, and when a prospect can quickly find a story that looks like their own. Put the result in the headline. Structure the story as Problem-Solution-Result, the format that Semrush, HubSpot and CoSchedule recommend for case studies, and keep it between 500 and 1,500 words. Eye-tracking research from CXL and Nielsen Norman Group supports the trust signals.

Case studies are worth the effort. They're rated the most effective mid- and bottom-funnel B2B content type (CoSchedule, citing MarketingCharts), and 82% of B2B marketers use them (Content Marketing Institute, 2015). When someone is close to buying, this is the page they want to read.

What structure should a case study follow?

Use Problem-Solution-Result (PSR), the structure recommended by Semrush, HubSpot and CoSchedule. In order:

1. Title: client name plus the key result
2. At a glance: two to four sentences and a few bullet stats
3. The challenge: the problem, with enough tension that a reader recognises their own pain
4. The solution: what you did, focused on one or two key moves
5. The results: the numbers, with context
6. A quote: the client, in their own words
7. A call to action

PSR should still read as a story. The challenge builds empathy, the solution is the turning point, and the results pay it off. No published A/B test shows PSR beating a pure narrative or an interview format. That's the field's biggest research gap. The expert advice all points the same way, though, so we treat PSR as the safe default for almost every business.

Should the result go in the headline?

Yes. The best B2B case studies lead with a specific, quantified result: "How Domino's decreased cost per acquisition by 65%" (Segment/Twilio), or "70% cut in productivity-tool costs" (Notion). The formula is client name + verb + specific metric + with what.

A precise number such as "460% increase in mobile traffic" reads as measured, which makes it more believable than "a significant traffic increase". Put the result in the headline and again in an "at a glance" box near the top. As HubSpot puts it, readers shouldn't have to reach the end to learn what happened. It's the inverted pyramid: conclusion first, then the story for anyone who wants it.

Why do faces beat logos?

People remember people. In CXL's eye-tracking study of around 200 participants, testimonials with photos were the most memorable form of social proof. Recall was significantly higher than for testimonials without photos (p=0.0035). Logos and numbers weren't memorable.

The photo has to be real, though. Nielsen Norman Group's eye-tracking research found that people scrutinise genuine portraits and largely ignore stock or decorative photos. In one tested example, a page with a generic stock photo received zero fixations.

In CXL's eye-tracking study of ~200 participants, testimonials with photos had significantly higher recall than those without (p=0.0035).

Our rule: the client's logo goes on the index card, and a real face goes beside the pull quote inside the story. If the client won't share a photo, attribute the quote with a full name and job title at minimum.

How long should a case study be?

HubSpot recommends 500 to 1,500 words for most businesses. Match the length to the complexity of the work:

• Snapshot, 300–500 words: e-commerce, social cards, early-stage companies with limited data
• Standard, 500–1,500 words: most service and SaaS businesses, and the safe default
• In-depth, 1,500–2,500 words: enterprise deals, agencies showing methodology, professional services

No study has pinned down an optimal length for case studies specifically, so treat these ranges as guidelines. Short stories packed with metrics suit e-commerce. Longer narratives suit agencies and consultancies, where the work itself is the product.

Will a case study bring in Google traffic?

Mostly not directly. An Ahrefs analysis of HubSpot, which had 8.2 million monthly organic visits, found no case studies among its top organic pages. They rank for low-volume, branded queries like "[brand] case study."

The SEO value is indirect. Case studies are a strong E-E-A-T signal, because Google rewards demonstrated real-world experience. They also create contextual internal links between your service pages and your proof. A case study can rank when its title targets a problem keyword, such as "How [Client] reduced churn by 40%", rather than just "[Client] Case Study."

Nielsen Norman Group's B2B usability testing across 179 sites with 79 participants found a task-success rate of only 58%. Case studies hidden behind registration walls were a common failure in that testing. Make them easy to find from your service pages.

How many case studies do you need?

Three to five is the credible minimum (practitioner consensus). Fewer than three looks cherry-picked. Beyond five, more case studies help when they cover new industries or use cases. A business serving four industries with three use cases might want eight to twelve over time. One well-produced story with a named client, real photos and specific metrics does more than ten vague "success stories."

FAQ

What's the single most important element of a case study?

A specific number in the headline, such as Domino's 65% decrease in cost per acquisition. Precision signals that the result was measured, and measured results read as credible. The best B2B case studies lead with one.

Should I use a video testimonial or a written one?

Both have a role. 85% of consumers say a video convinced them to buy (Wyzowl, 2026, n=266), but written testimonials with photos hold attention longer on the page (CXL). Default to a written quote with a photo on every page, and add a short video (under three minutes) when budget allows. Poor video quality hurts trust.

Is it better to name the client or keep them anonymous?

Name them whenever you can. SaaS companies name clients almost universally, and anonymising is mostly a professional-services habit where confidentiality demands it. A named client with a face is stronger proof, and there's no controlled data showing that anonymous case studies convert better.

What should the headline formula be?

Client name + verb + specific metric + with what. For example: "[Client] increases coupon send rate by 76% with [Product]." Keep it to roughly 60–80 characters so it displays cleanly in search.

The customer-stories pages we build put the headline number and the client's face where nobody can miss them. Bring the case study you're least happy with to a call.